IBM shares updates on DOE's Summit supercomputer With expectations that it will be the world's fastest and most powerful supercomputer, Summit is expected to be completed in early 2018.

Three years after the Department of Energy (DOE) tapped IBM to build its next generation of supercomputers, the new "Summit" supercomputer is on track to be completed and operational in early 2018, IBM shared Monday at the SuperComputing 2017 conference in Denver.

It's currently under development at the Oak Ridge National Laboratory (ORNL) in Tennessee:

The Oak Ridge Leadership Computing Facility's Summit supercomputer installation continues. Summit will deliver more than five times the computational performance of Titan's 18,688 nodes, using only approximately 4,600 nodes when it arrives in 2018.

Based on the IBM POWER9 architecture, the DOE expects Summit to be the world's fastest and most powerful supercomputer. IBM is working with Nvidia and Mellanox to build out Summit. The DOE has partnered with these companies and others in its race against China to reach exascale computing. In recent years, a pair of Chinese supercomputers have held the top two spots in the Top500 supercomputer list. The Titan system at ORNL ranks farther down.

The Oak Ridge Leadership Computing Facility's Summit supercomputer installation continues with the Mellanox team installing and wiring interconnects. To provide a high rate of I/O throughput, Summit's nodes will be connected in a non-blocking fat-tree using a dual-rail Mellanox EDR InfiniBand interconnect.

"The infrastructure is now complete and we're beginning to deploy the IBM POWER9 compute nodes," said Buddy Bland, Oak Ridge Leadership Computing Facility Director, regarding Summit in an IBM blog post. "The advanced capabilities of the IBM POWER9 CPUs coupled with the NVIDIA Volta GPUs will significantly advance the computational performance of DOE's mission critical applications."

OLCF-4 Summit supercomputer infrastructure installation, March 28, 2017.

Along with Summit, the Sierra system at Lawrence Livermore National Labs (LLNL) is also on track for completion by early next year. Both systems are expected to increase application performance by five to 10 times in comparison to their predecessors, the Titan at ORNL and Sequoia at LLNL, according to Dave Turek, IBM's VP of HPC IBM Systems.